This article describes a basic 1.5-T hepatic magnetic resonance (MR) imaging protocol, strategies for optimizing pulse sequences while managing artifacts, the proper timing of postgadolinium 3-dimensional gradient echo sequences, and an effective order of performing pulse sequences with the goal of creating an efficient and high-quality hepatic MR imaging examination. The authors have implemented this general approach on General Electric, Philips, and Siemens clinical scanners.
Key points
- •
The foundation for hepatic magnetic resonance imaging (MRI) includes T1-weighted images (including chemical shift imaging), T2-weighted images, fat suppression, and in most cases, contrast-enhanced images. Complementary techniques include balanced steady-state free precession and diffusion-weighted imaging.
- •
To maximize the clinical utility of hepatic MRI exams, each pulse sequence must be optimized while minimizing artifacts that interfere with interpretation.
- •
An understanding of the different types of gadolinium-based contrast agents (GBCAs) and the most important characteristics of each agent is needed to improve the diagnostic yield of the hepatic MRI exam.
- •
T1-weighted fat-suppressed gradient echo (GRE) sequences must be properly timed to account for the type of GBCA used while adjusting imaging parameters to maximize image quality.
- •
Most pulse sequences can be effectively performed after administering gadolinium. Exceptions include dual GRE sequences, single-shot fast-spin echo heavily T2-weighted sequences, high-resolution 3-dimensional MR cholangiopancreatography sequences obtained after gadoxetate disodium administration, and short TI inversion recovery sequences obtained after administration of extracellular space contrast agents.
Introduction
Compared with other hepatic imaging modalities including ultrasonography, contrast-enhanced ultrasonography, computed tomography (CT), and positron emission tomography-CT, magnetic resonance imaging (MRI) offers more comprehensive evaluation of the liver, establishing in many cases an accurate tissue diagnosis. To fully harness the power of MRI, the techniques must be optimized while minimizing artifacts interfering with interpretation. The foundation for hepatic MRI includes T1-weighted images (including chemical shift imaging), T2-weighted images, fat suppression, and in most cases, contrast-enhanced images. Complementary imaging sequences include balanced steady state free precession (BSSFP) and diffusion weighted imaging (DWI). Pregadolinium and postgadolinium fat-suppressed T1-weighted 3D gradient echo (GRE) sequences are generally the workhorse of the examination and must be properly timed to account for the type of gadolinium based contrast agent (GBCA) while adjusting imaging parameters to maximize image quality. Finally, certain pulse sequences can be performed after gadolinium administration to improve examination efficiency while maximizing the diversity of pulse sequences.
This article describes a basic 1.5-T hepatic MRI protocol, strategies for optimizing pulse sequences while managing artifacts, the proper timing of postgadolinium 3D GRE sequences, and an effective order of performing pulse sequences with the goal of creating an efficient and high-quality hepatic MRI examination. The authors have implemented this general approach on Philips (Philips Medical Systems, Best, Netherlands), Siemens (Siemens Medical Solutions, Erlangen, Germany) and General Electric (GE Medical Systems, Milwaukee, Wisconsin, USA) clinical scanners.
Introduction
Compared with other hepatic imaging modalities including ultrasonography, contrast-enhanced ultrasonography, computed tomography (CT), and positron emission tomography-CT, magnetic resonance imaging (MRI) offers more comprehensive evaluation of the liver, establishing in many cases an accurate tissue diagnosis. To fully harness the power of MRI, the techniques must be optimized while minimizing artifacts interfering with interpretation. The foundation for hepatic MRI includes T1-weighted images (including chemical shift imaging), T2-weighted images, fat suppression, and in most cases, contrast-enhanced images. Complementary imaging sequences include balanced steady state free precession (BSSFP) and diffusion weighted imaging (DWI). Pregadolinium and postgadolinium fat-suppressed T1-weighted 3D gradient echo (GRE) sequences are generally the workhorse of the examination and must be properly timed to account for the type of gadolinium based contrast agent (GBCA) while adjusting imaging parameters to maximize image quality. Finally, certain pulse sequences can be performed after gadolinium administration to improve examination efficiency while maximizing the diversity of pulse sequences.
This article describes a basic 1.5-T hepatic MRI protocol, strategies for optimizing pulse sequences while managing artifacts, the proper timing of postgadolinium 3D GRE sequences, and an effective order of performing pulse sequences with the goal of creating an efficient and high-quality hepatic MRI examination. The authors have implemented this general approach on Philips (Philips Medical Systems, Best, Netherlands), Siemens (Siemens Medical Solutions, Erlangen, Germany) and General Electric (GE Medical Systems, Milwaukee, Wisconsin, USA) clinical scanners.
The basic hepatic MR imaging examination
In clinical practice, the typical hepatic MR imaging examination ( Table 1 ) includes comprehensive imaging of other abdominal viscera, although generally, the MR imaging protocol is optimized for evaluation of the liver. For gadolinium-enhanced studies, intravenous gadolinium should be administered as early as possible during the examination. Then, if the examination is prematurely terminated for any reason, gadolinium-enhanced images, which are often the most important sequences for lesion characterization, will have already been completed. This protocol is achievable because most sequences, except dual GRE (in and out of phase) sequences, single shot fast spin echo (SSFSE) heavily T2-weighted sequences, high-resolution 3D MR cholangiopancreatography (MRCP) sequences obtained after gadoxetate disodium administration, and short TI inversion recovery (STIR) sequences obtained after administration of extracellular space contrast agents (ECSAs), are not adversely affected by gadolinium and can be performed after gadolinium administration. A torso phased array coil should be used for all sequences, including localizer images.
| Sequence | Plane | FOV | Matrix | PAF | Slice Thickness/Gap | TR (ms) | TE (ms) | Flip Angle | Fat Saturation |
|---|---|---|---|---|---|---|---|---|---|
| Protocol when using an extracellular space agent (ECSA): | |||||||||
| SSFSE survey | 3 plane | 48 | 320 × 192 | 0 | 8/0 | Min | 80 | 90 | No |
| SSFSE (Heavily T2WI) | Coronal | 44 | 256 × 192 | 1.7 | 5/0 | Min | 180 | 90 | No |
| SSFSE (heavily T2WI) | Axial | 38 | 256 × 192 | 2 | 5/0 | Min | 180 | 90 | No |
| Dual GRE in and out of phase | Axial | 38 | 256 × 192 | 2 | 7/0.5 | 265 | 2.1/4.4 | 90 | No |
| T1-weighted 3D spoiled GRE (pre-contrast double arterial, & portal venous phase) | Axial | 42 | 320 × 224 | 1.8 | 4.4/2.2 | Min | Min | 12 | Yes |
| Moderately T2W FS | Axial | 40 | 256 × 192 | 2 | 7.5/0.5 | 2300 | 84 | 90 | Yes |
| T1-weighted 3D spoiled GRE (delayed phase a ) | Axial | 42 | 320 × 224 | 1.8 | 4.4/2.2 | Min | Min | 15 | Yes |
| BSSFP | Axial | 38 | 192 × 288 | 2 | 5/0 | Min | Min | 70 | Yes |
| Radial Slab 2D MRCP b | Coronal | 26 | 288 × 256 | 0 | 40/0 | 2666 | 1096 | 90 | Yes |
| Diffusion | Axial | 36 | 128 × 160 | 2 | 6/1 | 7000 | 73 | 90 | Yes |
| 3D MRCP c | Coronal | 38 | 320 × 320 | 2 | 1.4/0.7 | 3750 | 847 | 90 | Yes |
| Additional postgadolinium pulse sequences and parameters when using gadoxetate disodium instead of an ECSA: | |||||||||
| T1-weighted 3D spoiled GRE (hepatobiliary phase) | Axial | 40 | 288 × 160 | 1.8 | 5/2.5 | Min | Min | 25–30 d | Yes |
| T1-weighted 3D spoiled GRE (hepatobiliary phase) | Coronal | 42 | 288 × 160 | 0 | 5/2.5 | Min | Min | 25–30 d | Yes |
a Delayed phase is called late dynamic phase when using gadoxetate disodium.
b Radial slab 2D MRCP can be performed if it is completed within 5 minutes of gadoxetate disodium injection.
c 3D MRCP should be performed before gadoxetate disodium administration or not at all.
d On some MR imaging scanners, the maximal flip angle for some 3D GRE pulse sequences is limited to 15° or so, in which case alternative pulse sequences can be considered to accommodate higher flip angles.
Individual pulse sequences
Localizer Images
Clinical utility
This series is used to confirm positioning of the torso coil by ensuring that the entire liver and other tissues of interest are included within the sensitive volume of the coil. Sagittal images are useful for determining the anteroposterior dimensions of the abdomen to localize subsequent axial images and help determine if a rectangular phase field of view can be used. Although T1-weighted images are commonly used for localizer images, a T2-weighted SSFSE sequence or T2/T1-weighted BSSFP sequence can be more clinically useful. The latter can be implemented as a rapid motion-insensitive 3-plane comprehensive initial survey.
The sagittal and/or coronal images obtained in a localizer sequence are useful for evaluating the spine to exclude compression fractures and to localize findings noted in the axial plane. A BSSFP localizer series is useful to evaluate vascular anatomy and patency (discussed below).
Technique/Optimization
The localizer series is usually a rapid sequence with a moderately large field of view obtained in 3 planes (ie, axial, sagittal, and coronal) or just the coronal plane. There are often significant gaps between slices to ensure adequate coverage in minimal time, but the authors recommend increasing the utility of these images by using a slice thickness of 6 mm or less and a gap between slices of 1 mm or less. A comprehensive 3-plane BSSFP survey of the entire abdomen can be completed in about 2 minutes during quiet breathing.
Artifacts
Large gaps between each slice, and variable breathing or breath-hold, cause misregistration artifact. Artifacts for SSFSE T2-weighted and BSSFP sequences are discussed below.
SSFSE Heavily T2-Weighted Images
Clinical utility
This fluid-sensitive series is useful for demonstrating cysts, fluid collections, or edema. For cystic lesions, this sequence provides superior visualization of papillary projections or septations compared with CT. In general, the vast majority of liver lesions with hyperintensity approaching cerebrospinal fluid on heavily T2-weighted images are benign (cysts or hemangiomas) and largely contain free water, whereas solid lesions lacking free water and generally composed of water bound to macromolecules are usually not well seen on this sequence. This fact is in contrast to moderately T2-weighted images (discussed below) in which both benign and solid lesions are generally conspicuous ( Fig. 1 ). The necessarily long echo time (TE) in the heavily T2-weighted sequence lends itself to the single-shot technique in which a single excitation pulse is followed by a long series of alternating 180° refocusing pulses and echoes accommodating the long TE. The multiplicity of refocusing pulses corrects for susceptibility artifact, and the SSFSE sequence is generally more resistant to this artifact than other sequences, especially GRE sequences.
Technique/Optimization
This series is relatively motion resistant with an acquisition duration of 1 second per slice or less. Because the acquisition time is relatively short, the sequence should be obtained during breath-holding. If the sequence has to be divided into 2 breath-holds to cover the entire examination volume, it should be acquired in a sequential manner (ie, images 1, 2, 3, etc.), not concatenated (ie, images 1, 3, 5, etc.), to reduce misregistration artifact. Then, if the patient has a different degree of inspiration between the 2 breath-holds, the scan will not seem to move back and forth on every other image when reconstructed. A TE greater than 160 milliseconds is considered heavily T2 weighted and helps differentiate solid from nonsolid lesions, although ideally the TE should be 180–200 milliseconds. Fat saturation is not recommended, because this obscures the clarity of liver and other visceral margins and further decreases overall signal, which is already relatively low (because of the single-echo acquisition technique). The authors recommend performing this series before contrast, which more reliably shows high signal intensity of cavernous hemangiomas and allows evaluation of the renal collecting systems without the T2-shortening effects of excreted gadolinium ( Fig. 2 ).
Artifacts
Fluid motion induces phase incoherence, which can lead to signal voids in ascites, and pleural effusions that can mimic disease ( Figs. 3 and 4 ) and make pleural and pericardial effusions less visible ( Fig. 5 ). The BSSFP series is less vulnerable to this artifact and supplements the SSFSE sequence in characterizing fluid collections (see Fig. 5 ). If this sequence is concatenated or split into more than one acquisition, misregistration artifact between adjacent slices occurs if the patient is breathing during image acquisition or uses variable breath-holding; this potentially leads to incomplete anatomic coverage, precluding evaluation of cystic lesions or biliary abnormalities. Non-fat-suppressed heavily T2-weighted images show crisp fluid margins, but soft tissue detail is generally blurry compared with 3D GRE sequences. Blurring is the consequence of the substantial T2 relaxation that occurs in tissues with short or moderate T2 relaxation times between the acquisition of the early and late echoes of the long echo train of the SSFSE sequence. The blurring of SSFSE images can be decreased by using parallel imaging.
Balanced Steady State Free Precession (BSSFP)
Clinical utility
BSSFP can be used as a rapid motion-insensitive localization sequence (discussed above) or as a stand-alone diagnostic sequence. On this series, both stationary fluid and flowing blood are depicted as bright, providing an effective survey of vascular and ductal structures, bowel, and fluid collections. This bright blood technique is especially useful in hepatic imaging when performing studies without contrast or when using gadoxetate disodium, in which case blood vessel enhancement may be less than optimal (discussed below). Although balanced SSFP images are less specific for identifying fluid and ductal structures compared with heavily T2-weighted SSFSE images, they are obtained more rapidly, have higher signal-to-noise ratio (SNR), and depict blood vessels better.
Specific vendor names for BSSFP sequences include true fast imaging with steady-state precession (true FISP), fast imaging employing steady-state acquisition (FIESTA), and balanced fast-field echo (BFFE).
Technique/Optimization
In a BSSFP sequence, magnetization is continually preserved rather than spoiled and each radio-frequency (RF) pulse excites as well as refocuses. By setting the repetition time (TR) below tissue T2 relaxation times, residual transverse and longitudinal magnetization is continuously flipped by successive excitation pulses, contributing to subsequent echoes. At some point, the magnitudes of the transverse and longitudinal magnetization stabilize—or reach a steady state. The greater the flip angle, the more transverse magnetization will be converted to longitudinal magnetization; typical flip angles for steady-state sequences are between 50° and 80°. The pulse sequences have a net phase change for all 3 gradients of zero. With a combination of T2* and T1 weighting, fat is bright on this sequence, potentially interfering with detecting fluid collections such as pericardial fluid. For this reason, performing fat saturation generally enhances the diagnostic yield, although this fat suppression may be poor at air-fat interfaces because of magnetic field inhomogeneity.
Artifacts
For the BSSFP localizer series, banding artifact can occur, because of incomplete refocusing in areas with magnetic field distortion ( Fig. 6 ). Gaps between RF pulses lead to incomplete refocusing and significant susceptibility artifact. Banding and susceptibility artifacts are both reduced by using the shortest possible TR.
Dual Gradient Echo Axial In-Phase and Out-Of-Phase (Dual Fast Field Echo)
Clinical utility
This series is useful for tissue characterization on multiple levels by providing sensitivity to microscopic fat, susceptibility artifact, and substances with short T1 values. The fat-water phase cancellation that occurs on the out-of-phase series allows this series to be used to detect microscopic fat in the liver characteristic of hepatic steatosis ( Fig. 7 ) and fat within liver lesions such as hepatocellular carcinoma or hepatic adenoma. The T2* effects of the longer TE on the in-phase images make this series sensitive for magnetic susceptibility. For this reason, this series is also clinically useful for detecting iron in abdominal organs (ie, hemosiderosis or hemochromatosis), calcium (ie, granulomas or biliary stones), air (ie, pneumobilia, portal venous air, bowel gas, or pneumoperitoneum), or metal objects such as surgical clips from prior surgery, embolization coils, or stents. This series is usually motion sensitive and identifies patient’s breath-holding limitations before performing gadolinium-enhanced 3D GRE sequences.
